I

O

PmR

AS

a

b

c

d

e

f

RA

h1

nternational Journal of Clinical and Health Psychology (2015) 15, 8---17

International Journalof Clinical and Health Psychology

www.elsevier.es/ijchp

RIGINAL ARTICLE

ersonal value orientations as mediated predictors ofental health: A three-culture study of Chinese,ussian, and German university students

ndreas Maerckera,∗, Xiao Chi Zhangb, Zhihua Gaoc, Yakov Kochetkovd,han Lue, Zhiqin Sangf, Shaoqing Yangc, Silvia Schneiderb, Jürgen Margrafb

University of Zurich, SwitzerlandRuhr University Bochum, GermanyHebei United University, ChinaMoscow Research Institute of Psychiatry, RussiaCapital Normal University, ChinaNanjing University, China

eceived 5 June 2014; accepted 23 June 2014vailable online 13 August 2014

KEYWORDSPersonal values;Value orientations;Mental health;Cross-culturalpsychology;Descriptive surveystudy

Abstract Previous studies of traditional and modern value orientations in individuals foundmediated predictive relationships of these values on particular mental disorders. The aim ofthis study with samples from three countries (Germany, Russia, and China) is to extend find-ings on mental health (MH) and value orientations to broader MH indicators and two typesof mediators, i.e. social support and resilience in accordance to a theory of values and mod-ernization/postmodernization. The multisite study was conducted in the three countries. Apath-model with traditional values predicting MH mediated by social support, and modernvalues predicting MH mediated by resilience was tested in all three countries. As expected,value orientations were for the most part strongest in China, followed by Russia and Germany.Structural equation modeling supported the assumption of mediated prediction of MH by valueorientations by and large. The traditional value benevolence predicts social support whereasthe modern value self-direction predicts resilience. Value orientations are a sensitive tool toempirically describe cross-cultural differences. The findings indicate that personal value ori-entations are meaningful predictors of MH. The analysis of personal values shows promise inlinking public health, cross-cultural and modernization issues.

© 2014 Asociación Espanola de Psicología Conductual. Published by Elsevier España, S.L.U.

∗ Corresponding author: Department of Psychology, University of Zurich, Binzmuhlestr 14/17, CH-8050 Zurich, Switzerland.E-mail address: maercker@psychologie.uzh.ch (A. Maercker).

ttp://dx.doi.org/10.1016/j.ijchp.2014.06.001697-2600/© 2014 Asociación Espanola de Psicología Conductual. Published by Elsevier España, S.L.U.

Open access under CC BY-NC-ND license.

Open access under CC BY-NC-ND license.

Personal value orientations as mediated predictors of mental health 9

PALABRAS CLAVEValores personales;Orientaciones devalor;Salud mental;Psicologíatranscultural;Estudio descriptivode poblaciones

Orientaciones de valor personal como predictores mediadores de la salud mental: unestudio transcultural con estudiantes universitarios de China, Rusia y Alemania

Resumen Estudios previos sobre orientaciones de valores tradicionales y modernos encon-traron relaciones predictivas mediadas con trastornos mentales particulares. El objetivo deeste estudio con muestras procedentes de tres países (Alemania, Rusia y China) es ampliar losresultados indicadores más amplios de salud mental (SM) y dos tipos de mediadores (apoyo socialy resiliencia), de acuerdo a la teoría de los valores y la modernización/postmodernización. Seprobó en los tres países un path-model con valores tradicionales prediciendo la SM mediadapor el apoyo social y valores modernos prediciendo la SM mediada por la resiliencia. Como erade esperar, las orientaciones de valores eran en su mayor parte más fuertes en China, seguidopor Rusia y Alemania. Modelos de ecuaciones estructurales apoyaron la hipótesis de la predic-ción de la SM mediada por las orientaciones de valores en general. La benevolencia de valorestradicionales predice el apoyo social, mientras que el valor autodirección moderna predice laresiliencia. Las orientaciones de valores constituyen una herramienta sensible para describirempíricamente las diferencias interculturales. Los resultados indican que las orientaciones devalores personales son predictores significativos de la SM. El análisis de los valores personalesse muestra prometedor en la salud pública transcultural y cuestiones de modernización.© 2014 Asociación Espanola de Psicología Conductual. Publicado por Elsevier España, S.L.U.

GqalW

mttsftptar

(22tsi

iicbppccB

Open access under CC BY-NC-ND license.

Basic value orientations recently received increasinginterest in studying occupational health issues (e.g., DiazBretones & Gonzalez, 2011; Wang, Lu, & Lu, 2014), healthconsequences of migrants (Vecchione, Caprara, Schoen,Castro, & Schwartz, 2012) or mental health in militaryservice (Zimmermann et al., 2014). Basic or cultural valuesrefer to things that individuals believe are worth desiringin life and thus guide their behavior. According to Rokeach(1973), value orientations include cognitive, affective, andbehavioral components; value orientations are cognitions ofthe desirable that foster emotions and motivate behavior.In the present study, value orientations are examined usingSchwartz’s model (1992), which was developed based on thework of Rokeach. Schwartz views value orientations as thecriteria people use to select and justify actions and to evalu-ate people and events. Value orientations thus serve as one’sown guidance of behavior and the evaluation of behavior ofothers.

Maercker et al. (2009) proposed to group the ten sin-gle value orientations by Schwartz into traditional (security,conformity, tradition, benevolence, universalism) versusmodern (self-direction, stimulation, hedonism, achieve-ment, power) values and to introduce both as basiccultural value orientations in mental health research.Traditional cultural values stress collectivism, submissiveself-restriction, preservation of traditional practices, pro-tection, and stability. Modern cultural values representmotivations to pursue personal success and dominance overothers or gratification for oneself. For trauma and stress-related disorders it has been shown that modern valuesare protective against posttraumatic stress via mediatingsocial sharing processes, while traditional values increasethis stress via the same mediation pathway (Maercker

et al., 2009). This contradicted the popular long-standingassumption that traditional values are better for healthpreservation while modern values derogate health (cf.

ac2

raham, 2010; Kleinman & Good, 1985). However, a subse-uent study suggested that these effect patterns are partlyffected when other types of health outcome and particu-ar age groups of adults are concerned (Müller, Forstmeier,agner, & Maercker, 2011).

The present large-scale multi-national survey of studentental health and its predictors allows broadening the scope

owards broader indicators of mental health, namely posi-ive mental health and depression. At the same time, theurvey study allows the study of broader health-promotingactors of social support and resilience as potential media-ors of the cultural value----mental health relationship. It isart of a larger research program investigating predictors,ranscultural aspects and time course of mental health in

series of cross-sectional and longitudinal studies (BOOMesearch program; Margraf & Schneider, 2014).

Based on theories of social belongingness by Baumeister2005) and the role of autonomy in resilience (Richardson,002), as well as own previous findings (Maercker et al.,009) we hypothesize a prediction model in which tradi-ional values mainly take effect on mental health via socialupport while modern values do so via the individuals’ capac-ty for resilience.

According to Baumeister (2005), all human beings aren need of a certain minimum quantity of regular, satisfy-ng social interactions. He argues that traditional humanulture is compelled and conditioned by pressure toelong and that this belongingness is relevant for theromotion of psychological wellbeing. This has been sup-orted in various lines of research (e.g., depression inhildren, caregivers of demented individuals, psychoedu-ational health interventions) (for a review see Kawachi &erkman, 2001). However, it has been shown that from the

ctor’s perspective, social ties may also entail psychologi-al costs in the sense of indebtedness and obligation (Cohen,004).

1

cdiaa2ftdrdma

otkowv(tms(tascets(tC

Rfuisvar

M

P

Qt(saVctoit

pcCplaqut

nbCwPaoamaTdifp

gatdf

M

FZbCtafrt‘i

sctddtlepis

0

Resilience is still a relatively recent construct in psy-hology with slightly divergent definitions. In some of theseefinitions it is congruently viewed as a construct, whichncludes variables like temperament and personality, inddition to specific skills (e.g. active problem-solving) thatllow individuals to cope well with life stress (Richardson,002). Following this, resilience refers to the motivationalorces of a person and his/her creation of experienceshat foster activation and utilization of own forces thatrive toward self-actualization and to reintegrate from dis-uptions. The conceptual closeness of resilience (in theseefinitions) to autonomy motivation leads us to assume thatodern values (e.g. self-direction, achievement, hedonism)

re predictors of resilience.So far, we have not discussed if different magnitudes

r patterns of traditional or modern values are assumedo have differential impacts on mental health. Commonnowledge suggests that there are differences in valuerientations across cultures or nations. A theoretical frame-ork capturing this assumption is Inglehart’s model ofalue persistence and modernization/postmodernizationInglehart, 1997; Inglehart & Baker, 2000). This model dis-inguishes different states of industrialization (concerningodernization) or the rise of the service and knowledge

ector (concerning post modernization) in society. Inglehart1997) proposes that traditional values are sustained duringhese processes or even strengthened while modern valuesre specific to modern and postmodern societies. This per-istence of traditional values despite economic and politicalhanges is a part of the cultural heritage of a given soci-ty. The model also argues that the differences betweenhe values held by members of different traditions in a givenociety are much smaller than are cross-national differencesInglehart & Baker, 2000, p. 18). Thus, a consequence of thisheory is that rapidly changing societies, like modern-dayhina, show both highest traditional and modern values.

The current study aims to investigate in a Chinese,ussian, and German sample of University students theollowing assumptions: (1) According to Inglehart’s val-es theory, traditional values should be most pronouncedn China, followed by Russia in contrast to lower repre-entation in Germany; (2) A path-model with traditionalalues predicting mental health mediated by social support,nd modern values predicting mental health mediated byesilience is valid in all three countries.

ethod

articipants

uestionnaires were administered in different ways in thehree countries. Data from students of universities in ChinaCapital Normal University Beijing, Hebei United Univer-ity, Nanjing University), Germany (Ruhr-University Bochum)nd Russia (Lomonossov University Moscow, University oforonesh, University of Orenburg) were gathered by theollaborating Departments of Psychology specifically for

he BOOM research program. In Germany, the Universityf Bochum set up an online portal for their students thatncluded the questionnaires. Data from this online inves-igation were available for N = 1.105 persons. In Russia,

ttcv

A. Maercker et al.

aper-and-pencil questionnaires were applied at the threeollaborating universities to N = 3.743 students. In the threehinese sites, from the N = 9,019 students, 59% partici-ated in grouping-online sessions that gathered in computerabs and were supervised by 1-2 research investigatorsnd 41% participated in paper-pencil (41%) testing. Alluestionnaires were administered in the native languagessing either available publishes versions or the customaryranslation-backtranslation method.

Since the data were anonymized from the very begin-ing of data collection, no statement by an institutionaloard/ethics committee was required for/in Russia andhina. The study in total and in specific its German partas approved by the ethics committee of the Faculty ofsychology at Ruhr-University on May 12, 2011. Approval todminister the questionnaires was granted by the Facultyf Psychology at Ruhr-University Bochum on May 12, 2011nd renewed on October 2012. The approvals for the Ger-an site were communicated to the participating Chinese

nd Russian Universities who acknowledged these approvals.he Chinese and Russian samples included University stu-ents below age 18. Chinese and Russian laws grant officiallynscribed University students of all ages the rights to decideor themselves about study-related issues including partici-ation in studies.

Table 1 shows the demographic characteristics. Whileender ratios are not different between the samples, meange was (F = 2949.07; df = 2; p < .001) with the German par-icipants being older than the Chinese or Russians. Otherifferences were marital status and the affluence of parent’samily.

easures

amily Affluence Scale. The FAS (Boyce, Torsheim, Currie, &ambon, 2006), a four-item measure of family wealth, haseen developed in the WHO Health Behavior in School-agedhildren Study and has been validated by correlations withhe Gross National Product across 35 countries. Its questionsre (either with 2 or 3 response alternatives): ‘Does youramily own a car, van or truck?’, ‘Do you have your own bed-oom for yourself?’, ‘During the past 12 months, how manyimes did you travel away on holiday with your family?’, andHow many computers does your family own?’. The FAS scores calculated by summing up the responses to these items.

Value orientations. Basic value orientations were mea-ured with the PVQ 21 item version that is based on theircumflex models of basic values and had been used inhe European Social Survey (Schwartz, 2007). Each itemescribes a person in two sentences (‘portrait’). Respon-ents are asked to assess how similar to the portrait personhey are. Answers range from ‘very similar’ to ‘very dissimi-ar’, coded from 1 to 6. Two items each measured the valuesxcept that three items measured universalism. For exam-le, two items measured hedonism (‘Having a good time ismportant to her. She likes to ‘‘spoil’’ herself.’, and ‘Sheeeks every chance she can to have fun. It is important to her

o do things that give her pleasure.’). In the current studywo sum scales were calculated (traditional values: security,onformity, tradition, benevolence, universalism; modernalues: self-direction, stimulation, hedonism, achievement,

Personal value orientations as mediated predictors of mental health 11

Table 1 Demographic features of the three samples.

Germany (N = 1,105) Russia (N = 3,743) China (N = 9,019)

GenderMales 39.7% 34.5% 36%Females 60.3% 65.5% 64%Age: Mean in years (SD) 25.23 (3.71) 19.82 (2.36) 19.85 (1.89)

Marital StatusSingle 33.4% 49.4% 81.9%in a relationship 60.7% 42.5% 17.3%Married 5.5% 6.7% 0.8%Divorced 0.4% 1.3% 0.0%

Family Affluence ScaleLow 6.4% 17.1% 48.7%Medium 47.8% 53.4% 36.5%

s2istr

S

Wwavo

tvmltdveortiii

ttmaga

High 45.8%

power). The native language versions of the PVQ have beenvalidated previously for the languages (see European SocialSurvey, ESS, 2008; Maercker et al., 2009). The alpha reli-ability coefficients for traditional/modern value sum scoreswere, respectively, .67/.73 in Germany; .76/.78 in Russia;.77/.77 in China. Reliability coefficients of the ten basic val-ues ranged from .32 (tradition/customs orientation in China)to .73 (achievement in Germany).

Resilience. The scale (Schumacher, Leppert, Gunzel-mann, Strau�, & Brähler, 2005; Wagnild & Young, 1993)consists of 11 items assessing personality characteristics(e.g., ‘I like me’, ‘I keep interested in many things’, ‘Ihave enough energy’) on a 7-point Likert scale. The Germanversion has been validated in various studies (Schumacheret al., 2005). The Chinese version was previously validated(Gao, Yang, Margraf, & Zhang, 2013). The alpha reliabil-ity coefficients were, respectively, .94 in Germany; .81 inRussia; .82 in China.

Social support. The F-SOZU is a well-established scalewith 14 items (Fydrich, Sommer, Tydecks, & Brähler., 2009).It measures perceived emotional and instrumental supportand social integration by items with a 5-point scale in onesum score (sample item: ‘‘I know whom I can go to when-ever I feel low’’). The alpha reliability coefficients were,respectively, .95 in Germany; .94 in Russia; .95 in China.

Positive mental health. The 9 items questionnaire (Lukat,Margraf, Becker, van der Veld, & Lutz, 2014) comprises state-ments like: ‘Much of what I do brings me joy’, ‘I enjoy mylife’, ‘I am often carefree and in good spirits. An earlierversion of the scale has been successfully used in our ear-lier Dresden Predictor Study (see Trumpf, Becker, Vriends,Meyer, & Margraf, 2009; Trumpf et al., 2010) where it showedgood reliability and predictive power for various anxiety dis-orders (Vriends, Becker, Meyer, Michael, & Margraf, 2007;Vriends, Becker, Meyer, Williams et al., 2007). No previousvalidation existed in Russian and Chinese language; the cus-tomary translation-backtranslation-modification procedurewas thus applied. The alpha reliability coefficients in the

present study were, respectively, .93 in Germany; .86 inRussia; .90 in China. To complement the salutogenic mentalhealth construct, the following more traditional depressionscale was applied in parallel.

fis

s

29.4% 14.8%

Depression was measured by a subscale of the Depres-ion Anxiety Stress Scales 21 (DASS-21; Henry & Crawford,005) which showed excellent psychometric properties ints validation study. The seven depression items are pre-ented with a 4-point Likert scale. Validated versions exist inhe three languages. The alpha reliability coefficients were,espectively, .89 in Germany; .83 in Russia; .79 in China.

tatistical analysis

e first compared means of all model variables. Second,e computed zero-order bivariate correlations to assess thessociation between model variables and the main outcomesariable of positive mental health. Sum scores for traditionalr modern values were additionally calculated.

We then used structural equation modeling. With regardo path analysis, from the five traditional and five modernalues in each case two single values were selected as basicodel predictors. Criteria were comparably highest corre-

ations with mental health and not being direct neighbors inhe circumflex model of values. Subsequently, the two tra-itional values predicted social support and the two modernalues predicted resilience. We allowed the single value ori-ntations to covariate freely as well as the error variancesf the factors. Reciprocal paths between social support andesilience were also allowed. In addition to positive men-al health as model outcome variable also depression wasncluded into the model, mainly because the latter measures more established and validated and thus allows compar-sons with already existing studies.

We assessed the comprehensive measurement model inhe same way in all three countries. We began by assessinghe measurement model separately in each country. We usedaximum likelihood to estimate parameters with SPSS Amos

nd evaluated the covariance structure models with severaloodness of fit indices: chi-square, root mean square error ofpproximation (RMSEA, Steiger & Lind, 1980), comparative

t index (CFI, Bentler, 1990), and standardized root meanquare residual (SRMR, Jöreskog & Sörbom, 1996).

Finally, we ran a multi-group analysis to assess mea-urement invariance of the scales of the variables across

12

A. M

aercker et

al.

Table 2 Pearson bivariate correlations of positive mental health with socio-demographic, value-orientation, and other model variables.

Means, SD, Tests Correlations with positivemental health (Pearson’s r)

Germany Russia China Tests (F, p, contrasts) Germany Russia China

Demographic variablesGender - - - - .08** .03 −.02*

Age 25.23(3.71) 19.82(2.36) 19.85(1.89) 2949.07; p < .001; R,C < G −.08** .04** −.01Family Affluence Scale 5.26(1.79) 4.37(1.91) 2.97(2.19) 1001.79; p < .001; C < R < G .12*** .16*** .04***

Value orientationsSecurity 8.10(2.03) 7.34(2.33) 8.65(2.01) 515.40; p < .001; R < G < C −.01 .09*** .07***

Conformity 7.01(2.31) 8.26(2.24) 9.20(1.93) 717.51; p < .001; G < R < C −.04 .12*** .19***

Tradition 7.49(2.06) 7.47(2.23) 8.18(2.06) 179.16; p < .001; G,R < C .02 .09*** .09***

Benevolence 10.25(1.49) 9.26(1.91) 10.08(1.67) 332.15; p < .001; R < C < G .14*** .26*** .35***

Universalism 14.32(2.51) 12.67(2.75) 14.72(2.41) 876.69; p < .001; R < G < C .09** .16*** .34***

Self-direction 9.21(1.78) 8.34(2.09) 8.41(2.01) 85.00; p < .001; R,C < G .24*** .22*** .28***

Stimulation 7.17(2.30) 7.99(2.28) 8.35(2.34) 138.94; p < .001; G < R < C .27*** .21*** .20***

Hedonism 8.76(2.10) 8.58(2.22) 9.50(2.00) 289.86; p < .001; R < G < C .45*** .29*** .33***

Achievement 8.73(2.10) 7.77(2.33) 9.11(2.01) 540.35; p < .001; R < G < C −.00 .13*** .20***

Power 7.01(2.01) 7.08(2.24) 7.44(2.05) 49.80; p < .001; G,R < C .03 .09*** -.05***

Tradit. values 47.18(6.41) 45.02(8.03) 50.8(7.14) 873.41; p < .001; R < G < C .05 .20*** .30***

Modern values 40.88(6.53) 39.76(7.83) 42.78(7.11) 243.09; p < .001; R < G < C .31*** .27*** .28***

Other model variablesPositive mental health 18.23(5.94) 18.99(5.08) 21.47(4.91) 444.83; p < .001; G < R < C - - -Social support 59.11(11.08) 58.00(11.21) 57.04(12.09) 20.53; p < .001; C < R < G .43*** .43*** .35***

Resilience 55.73(14.51) 59.28(8.70) 58.82(8.16) 68.97; p < .001; G < R < C .38*** .52*** .53***

Depression 4.41(4.39) 4.40(4.14) 1.66(2.34) 1219.45; p < .001; C < G,R -.72*** -.58*** -.52***

Note: R < G < C: Russian significantly lower than German group which in turn is significantly lower than Chinese Group; R,G < C: Russian and German Groups significantly lower than ChineseGroups; etc.

Personal value orientations as mediated predictors of mental health 13

Table 3 Goodness of fit statistics for the models tested.

Model fit indices Model comparison test

�2 df CFI SRMR RMSEA Comparison ��2 �df p

Measurement modelGermany 4080.70 1097 .91 .06 .05(.05-.05)Russia 7955.78 1097 .91 .06 .04(.04-.04)China 20760.86 1097 .90 .07 .04(.04-.05)

Tests of cross-national invarianceModel 1. Configural invariance 32798.36 3291 .90 .06 .02(.02–03)

.07

tidG

M

Aldt�(SsR

smialmC(ehmptppp

csptttm

Model 2. Full metric invariance 35238.73 3373 .90

country samples. Model 1 assessed configural invariance ---equivalence of the overall structure (i.e., the same num-ber of factors and the same pattern of loadings) acrosscountries. Model 2 examined metric invariance or equiv-alence of factor loadings across countries. These modelswere subsequently re-calculated for controlling for sampledifferences in gender, age, and family affluence by usingparallelized subsamples.

Results

Cross-national differences in values

Table 2 shows that four of five traditional values are highestin China with the exception of benevolence, which is high-est in Germany. These four values by and large consistentlyfollow the rank order of China (highest), Russia (medium)and Germany (lowest). Of the five modern values, four arealso highest in China with the exception of self-direction,which is highest in Germany. The rank order changed acrossthe values with lowest scores in Russia for self-direction,hedonism and achievement and lowest scores in Germanyfor stimulation and power. The comparisons of sum scores(traditional, modern), however, established a general rankorder: Russian lower than German lower than Chinese val-ues.

Predicting mental health

All model variables were single-order correlated (Pear-son’s r) with positive mental health (Table 2, last threecolumns). For Russia and China, all single value orientationswere significantly associated with positive mental healthon a low level (r < .30) with three exceptions of somewhathigher correlations only for China (r < .40: benevolence,universalism, hedonism). For Germany, a more differ-entiated pattern appeared: two modern (achievement,power) and three traditional values (security, conformity,tradition) were not related with mental health; hedo-nism was most pronouncedly associated (r = .46) whilethe remaining again only showed low associations. All

model variables had lower absolute correlation values withdepression compared to its correlations with well-beingexcept for rconformity-depression = .08 (p < .01) in Germany andrpower-depression = .12 (p < .001) in China.

bamp

.03(.03-.03) 2 vs. 1 2440.4 82 p < .001

Social support and resilience showed mid-sized correla-ions with positive mental health (from r = .35----.53). Thenverse association between positive mental health andepression ranked from China (r = -.53), Russia (r = -.59) toermany (r = -.73) (all p < .0001).

easurement models and tests for invariance

first model specified that conformity and benevo-ence exclusively predicted social support while self-irection and hedonism predicted resilience. Two ofhree country-specific models failed confirmation: Germany2(13) = 265.97, p < .001, CFI = .87, SRMR = .09, RMSEA = .13.12---.15) and China �2(13) = 1270.60, p < .001, CFI = .92,RMR = .07, RMSEA = .10 (.10---.11). Only the model for Rus-ia fitted: �2(13) = 373.62, p < .001, CFI = .94, SRMR = .05,MSEA = .09 (.08---.09).

After allowing hedonism additionally predicting socialupport and (directly) positive mental health, the threeodels for each country fit the data very well with all fit

ndices within acceptable ranges (see Table 3, upper part,nd Fig. 1). The factor loading of the indicators on theatent variables were for all relevant loadings in the Ger-an sample > .42; in the Russian sample > .10; and in thehinese sample > .15 and significantly different from zerop < .001). Some exceptions with non-significant loadingsmerged: self-direction with social support, positive mentalealth and depression; conformity with resilience, positiveental health and depression; benevolence with resilience,ositive mental health and, subsequently, depression. Thus,he model with its differentiation between social sup-ort and resilience predictors and mediated effects onositive mental health or, subsequently, depression was sup-orted.

Having confirmed the same measurement model in eachountry, we were able to test the invariance of mea-urement across countries. The fit statistics in the lowerart of Table 3 indicate an adequate (moderate) fit forhe configural invariance model (Model 1). This suggestshat the same factor structure holds for the three coun-ry samples. The fit statistics for the metric equivalenceodel (Model 2), which constrained all factor loadings to

e equal across groups, however, failed an adequate fit. Inddition, the chi-square difference test between the twoodels was statistically significant (delta �2 (81) = 2440.4,

< .001), supporting the non-equivalence so that metric

14 A. Maercker et al.

Germany

Conformity

Benevolence

Hedonism

Self direction

Resilience Depression

Positive mental health

Social support

e2

e1

e3

e4

.26

.21

.30

.15

–.05

.15 .32

.05 .10

.10

.25

.27

.37

–.73

.53

.15

.33

.24

–.15

–.10

Russia

Conformity

Benevolence

Hedonism

Self direction

Depression

Positive mental health

Resilience

Social support

e2

e1

e3

e4

.34

.12

.16.22

.32

.30.19

.05

.48.15

.41

.11

–.26

–.03

–.59

.28

.38

.34

.02

.19

.35

China

Conformity

Benevolence

Hedonism

Self direction

Depression

Positive mental health

Resilience

Social support

e2

e1

e3

e4

.40

.08

.45

.27

.00.05

.28

.66

.34

.28

.15–.53

–.53

–.19

.43

.17.23

.42.15

.36

.34

F na (cR ith lo

ito2l

tt

bnrn

igure 1 Final path models for Germany (a), Russia (b), Chiussian sample with loadings >.02 and in the Chinese sample w

nvariance could not be retained. The model did not meethe minimum requirement for metric invariance of at leastne equal loading for each measure (Meredith & Horn,001). Therefore only configural invariance was estab-

ished.

Finally, we repeated the same model estimations andests with subsamples that were parallelized for age, mari-al status and family affluence since these variables differed

wfi(s

). All paths in the German sample with loadings >.10, in theadings >.01 are significant.

etween the three national samples. For each subsample = 220 participants were included. By and large resultsemained identical with configural invariance confirmed butor metric invariance for all three national samples. CFIs

ere somewhat reduced but the other indices for misspeci-ed factor loadings (RMSEA) and for model misspecificationSRMR) remained acceptable indicating proof of the mea-urement model.

l he

fitp(hoTib2tut

tddwhtlpapchasrnoe

grfipsftaaampeiotmssFsso

Personal value orientations as mediated predictors of menta

Discussion

The study investigated if personal value orientations aremeaningful predictors for mental health outcomes. Itapplied a distinction into two value categories, tradi-tional and modern values (Ingelhart, 1997) that was alreadyapplied to previous studies on values and mental health(Maercker et al., 2009; Müller et al., 2011). Values orsocietal virtues are increasingly recognized as organiz-ing principles for human behaviors, cognitions and affectexpressions. Chinese virtues ( ) like ren (‘‘humanity’’),xiao (‘‘filial piety’’), and li (‘‘proper behavior’’), Ger-man virtues like Fleiss (‘‘diligence’’) or Pünktlichkeit(‘‘punctuality’’), and Russian virtues like(‘‘warm-heartedness’’) or (‘‘generosity’’) can beregarded as examples of such national traditional values.

First, we hypothesized according to Inglehart’s valuestheory (Inglehart & Baker, 2000), that traditional valueswould be most pronounced in China, followed by Russia incontrast to lower representation in Germany. Central partsof this hypothesis got supported with China having high-est extents of value orientations that supported Inglehart’stheory. It is also corroborated by findings by Jin, Zheng, &Xin (2009), who developed a Chinese Values Questionnaire(CVQ) and showed a tendency for the Chinese to think interms of the ‘‘Good-person’’, indicated by high scores onvalues of integrity, self-support, respect, and public inter-ests. However, the Russian student sample did not rankmidway but lowest for the value orientations. The early soci-ologist and social psychologist Émile Durkheim (1897/2009)had introduced the concept of ‘anomie’ to describe societalstates in which ‘disorientation’ and the lack or frailty ofvalue orientations is predominant, usually in countries thatexperienced setbacks. This concept could possibly explainwhy the value orientations in Russia by and large were low-est.

Exceptions of the hypothized pattern also appearedwith regard to single values of benevolence (a traditional)and self-direction (a modern value) that were the onlyones that are most pronounced in the German samplescompared to the other samples. Convergent evidence forthis finding exists only from research on temporal trendsin psychopathology during the last 80 years where forthe USA it was found that extrinsic goals including self-expression or narcissism increased while intrinsic goals suchas affiliation decreased (Twenge et al., 2010). The findingsfor benevolence may be corroborated by psycho-historicaldata that were collected by Pinker (2011) showing anincrease in benevolence emerging with economic progress(cf. Inglehart, 1997).

Second, we hypothesized that traditional values wouldpredict mental health mediated by social support,while modern values predict mental health mediated byresilience----and that this is valid in all three countries.We applied a simplified modeling that only included twosingle value orientations for each the traditional and mod-ern values (single values with the highest correlations withmental health across the three samples). By and large

the structural equation model supported the theoreticalmodel assumption with benevolence (significantly) and con-formity (non-significantly) predicting social support, andself-direction supporting resilience.

amcfi

alth 15

This partial support for the model extends previousndings (Maercker et al., 2009; Müller et al., 2011) onhe central role of value orientations for social-supportrocesses in disorders specifically associated with stressMaercker et al., 2013). In these studies traditional valuesad deteriorating effects on PTSD, but had relieving effectsn overcoming complicated grief or adjustment disorder.hus, the notion of a Janus phase of social support processes

n conjunction with mental health that had been outlinedy many authors (e.g., Cohen, 2004; Kawachi & Berkman,001) is again supported. Of particular importance remainshe overall finding of the present study that traditional val-es seem primarily important for social support with regardo mental health.

Regarding resilience as a mediator for mental health,he findings looked more mixed. Only self-direction pre-icted resilience as hypothesized according to a resilienceefinition that highlights the individual’s ability to copeell with life stress (Richardson, 2002). The modern valueedonism failed to directly predict resilience; instead it con-ributed directly to positive mental health and----somewhatess----social support. Hedonism is still an under-researchedhenomenon in psychology, although it seems relevant as

core process of self and identity regulation in contem-orary psychology (Baumeister, 2011). One obvious concernould be that the constructs of positive mental health andedonism overlap to explain the aberration of our hypothesist stake, but a view on the item formulations (see methodection above) does not suggest a complete overlap. In aecent study on German military service personnel, hedo-ism also stood out with the comparably highest predictionf all value orientations for mental health (Zimmermannt al., 2014).

There are several limitations to the present study. First,iven the cross-sectional nature of the data, conclusions oneal causality cannot be drawn. Although the path modelst the data, only experimental or longitudinal studies canrovide firm confirmation of causal or conditional relation-hips. Second, the three samples differ on demographiceatures, with the German sample being somewhat older;he German as well as Russian samples at higher percent-ges in a romantic relationship and from a higher familyffluence backgrounds. However, control analyses by par-llelized subsamples provided proof of our original pathodels albeit with lower test values. Furthermore, otherotential demographic confounders were not investigated,.g. rural or urban family background, which may play a rolen the processes under investigation. Third, the data werebtained exclusively by self-report. This method has cer-ain advantages for cross-cultural survey research, e.g. theeasures are economical and easy to administer). However,

elf-report measures may fail to capture well-being, depres-ion or value orientation processes in their full complexity.inally, generalization of the results towards other age orocietal groups than University students is not possible yet,ince the study focused only on those better-educated partsf the society.

In conclusion, the present study tries to contribute to

more culturally oriented research agenda for predictingental health in students by values and the main psychoso-

ial constructs of social support and resilience. Its mainndings of differential pathways of traditional or modern

1

vm

R

B

B

B

B

C

D

D

E

F

G

G

H

I

I

J

J

K

K

L

M

M

M

M

M

P

R

R

S

S

S

S

T

T

T

V

V

6

alues to mental health still needs corroboration and refine-ent through other research.

eferences

aumeister, R. F. (2005). The cultural animal: Human nature, mean-ing, and social life. Oxford: Oxford University Press.

aumeister, R. F. (2011). Self and identity: A brief overview of whatthey are, what they do, and how they work. Annals of the NewYork Academy of Sciences, 1234, 48---55. doi:10.1111/j.1749-6632.2011.06224.x.

entler, P. M. (1990). Comparative fit indexes in structural models.Psychological Bulletin, 107, 238---246.

oyce, W., Torsheim, T., Currie, C., & Zambon, A. (2006). The FamilyAffluence Scale as a measure of national wealth: Validation ofan adolescent self-report measure. Social Indicators Research,78. 473---487, doi:10.1007/s11205-005-1607-6.

ohen, S. (2004). Social relationships and health. American Psychol-ogist, 59, 676---684. doi:10.1037/0003-066X.59.8.676.

iaz Bretones, F., & Gonzalez, M. J. (2011). Subjective and occu-pational well-being in a sample of Mexican workers. SocialIndicators Research, 100, 273---285, doi:10.1007/s11205-010-9616-5.

urkeim, É. (1987/2009). Suicide: A study in Sociology. In M.Hechter, & C. Horne (Eds.), Theories of social order (pp.232---245). Stanford, CA: Stanford Social Sciences.

uropean Social Survey, ESS (2008). ESS4-2008 documenta-tion report [accessed February 7, 2014] From http://www.europeansocialsurvey.org/index.php?option=com content&view=article&id=65:supplementary-questionnaire&catid=116:questionnaire&Itemid=107

ydrich, T., Sommer, G., Tydecks, S., & Brähler, E. (2009). SocialSupport Questionnaire (F-SozU): Standardization of short form(K-14). Zeitsch Medizin Psychol, 18, 43---48.

ao, Z., Yang, S., Margraf, J., & Zhang, X. (2013). Reliability andValidity Test for Wagnild and Young’s Resilience Scale (RS-11) inChinese. China Journal of Health Psychology, 21, 1324---1326.

raham, G. (2010). The disordered mind: An introduction to phi-losophy of mind and mental illness. London: Routledge.

enry, J. D., & Crawford, J. R. (2005). The short-form versionof the Depression Anxiety Stress Scales (DASS-21): Constructvalidity and normative data in a large non-clinical sam-ple. British Journal of Clinical Psychology, 44, 227---239,doi:10.1348/014466505X29657.

nglehart, R. (1997). Modernization and postmodernization: Cul-tural, economic, and political change in 43 societies. Princeton,NJ: University Press.

nglehart, R., & Baker, W. E. (2000). Modernization, cultural change,and the persistence of traditional values. American SociologicalReview, 65, 19. doi:10.2307/2657288.

in, S., Zheng, J., & Xin, Z. (2009). The structure and characteristicsof contemporary Chinese values. Acta Psychologica Sinica, 41,1000---1014.

öreskog, K., & Sörbom, D. (1996). LISREL: 8 User’s ReferenceGuide. Chicago: Scientific Software International Inc.

awachi, I., & Berkman, L. F. (2001). Social Ties and Men-tal Health. Journal of Urban Health, 78, 458---467.doi:10.1093/jurban/78.3.458.

leinman, A., & Good, B. (1985). Culture and depression: Studiesin the anthropology and cross-cultural psychiatry of affect anddisorder. Berkeley, CA: University of California Press.

ukat, J., Margraf, J., Becker, E.S., van der Veld, W., & Lutz, R.(2014). Test criteria of the positive mental health scale (P-

scale). University of Bochum: Unpublished manuscript.

aercker, A., Brewin, C. R., Bryant, R. A., Cloitre, M., vanOmmeren, M., Jones, L. M., Humayan, A., Kagee, A., Llosa,A. E., Rousseau, C., Somasundaram, D. J., Souza, R., Suzuki,

V

A. Maercker et al.

Y., Weissbecker, I., Wessely, S. C., First, M. B., & Reed, G. M.(2013). Diagnosis and classification of disorders specifically asso-ciated with stress: Proposals for ICD-11. World Psychiatry, 12,198---206. doi:10.1002/wps.20057.

aercker, A., Mohiyeddini, C., Müller, M., Xie, W., Hui Yang, Z.,Wang, J., & Müller, J. (2009). Traditional versus modern values,self-perceived interpersonal factors, and posttraumatic stress inChinese and German crime victims. Psychology and Psychother-apy, 82, 219---232. doi:10.1348/147608308X380769.

argraf, J., & Schneider, S. (2014). Bochum Optimism and Men-tal Health (BOOM) Research Program: Background, methods andaims. Manuscript in preparation.

eredith, W., & Horn, J. (2001). The role of factorial invari-ance in modeling growth and change. In L. M. Collins, & A.G. Sayer (Eds.), New methods for the analysis of change.Decade of behavior (pp. 203---240). Washington, DC: Am PsycholAssoc.

üller, M., Forstmeier, S., Wagner, B., & Maercker, A. (2011).Traditional versus modern values and interpersonal factorspredicting stress response syndromes in a Swiss elderlypopulation. Psychology, Health & Medicine, 16, 631---640.doi:10.1080/13548506.2011.564192.

inker, S. (2011). The better angels of our nature: Why violence hasdeclined. New York: Penguin.

ichardson, G. E. (2002). The metatheory of resilienceand resiliency. Journal of Clinical Psychology, 58,307---321.

okeach, M. (1973). The nature of human values. New York, NY: TheFree Press.

chumacher, J., Leppert, K., Gunzelmann, T., Strauß, B., & Brähler,E. (2005). Resilience scale - a questionnaire to assess psycho-logical resilience as personality trait. Zeitschrift fur klinischePsychologie und Psychotherapie, 53, 16---39.

chwartz, S. H. (1992). Universals in the content and structure ofvalues: Theoretical advances and empirical tests in 20 countries.In M. P. Zanna (Ed.), Advances in Experimental Social Psychology(25) (pp. 1---65). Amsterdam: Elsevier.

chwartz, S. H. (2007). Universalism values and the inclusivenessof our moral universe. Journal of Cross-Cultural Psychology, 38,711---728.

teiger, J. H., & Lind, J. C. (1980). Statistically-based tests forthe number of factors. Iowa, IA: Annual Spring Meeting of thePsychometr Soc.

rumpf, J., Becker, E. S., Vriends, N., Meyer, A. H., & Margraf,J. (2009). Rates and predictors of remission in young womenwith specific phobia: A prospective community study. Journal ofAnxiety Disorders, 23, 958---964.

rumpf, J., Vriends, N., Meyer, A. H., Becker, E. S., Neumer,S., & Margraf, J. (2010). The Dresden Predictor Study ofAnxiety and Depression (DPS): Objectives, design, and meth-ods. Social Psychiatry and Psychiatric Epidemiology, 45,853---864.

wenge, J. M., Gentile, B., DeWall, C. N., Ma, D., Lacefield, K., &Schurtz, D. R. (2010). Birth cohort increases in psychopathol-ogy among young Americans, 1938-2007: A cross-temporalmeta-analysis of the MMPI. Clinical Psychology Review, 30,145---154.

ecchione, M., Caprara, G., Schoen, H., Castro, J. L., & Schwartz, S.H. (2012). The role of personal values and basic traits in percep-tions of the consequences of immigration: A three-nation study.British Journal of Psychology, 103, 359---377.

riends, N., Becker, E. S., Meyer, A., Michael, T., & Margraf, J.(2007). Subtypes of social phobias: Are they of any use? Journalof Anxiety Disorders, 21, 59---75.

riends, N., Becker, E. S., Meyer, A., Williams, L. S., Lutz, R., & Mar-

graf, J. (2007). Recovery from social phobia in community andits predictors: Data from a longitudinal epidemiological study.Journal of Anxiety Disorders, 21, 320---337.

l he

ZS., Backus, J., & Maercke, A. (2014). Personal values in sol-diers after military deployment: Associations with mental health

Personal value orientations as mediated predictors of menta

Wagnild, G. M., & Young, H. M. (1993). Development and psycho-metric evaluation of the Resilience Scale. Journal of NursingMeasurement, 1, 165---178.

Wang, H. J., Lu, C. Q., & Lu, L. (2014). Do people with traditionalvalues suffer more from job insecurity? The moderating effectsof traditionality. European Journal of Work & OrganizationalPsychology, 23, 107---117.

alth 17

immermann, P., Willmund, G. D., Kowalski, J., Firnkes, S., Siegel,

and resilience. European Journal of Psychotraumatology, 5. doi:10.3402/ejpt.v5.22939.